At the same
time, however, more competition also reduces bank risk. In a similar model, employing
the Panzar and Rosse H-statistic as a measure of competition, Schaeck, Cihak, and
Wolfe (2009) also find a negative relationship between the likelihood of a systemic banking
crisis and the competitiveness of the banking system. Another set of research uses (crosscountry)
bank-level data to examine the relationship between competition and individual
bank risk. Boyd, De Nicolo, and Jalal (2007) find that less-concentrated banking markets
are characterized by lower z-scores, an inverse measure of bank risk. Jiménez, Lopez, and
Saurina (2007) find no relationship between credit risk and market concentration but a
positive effect of competition on credit risk, supporting the competition-fragility hypothesis.
Schaeck and Cihak (2012) find a positive relationship between higher competition
and bank capital ratios. Inasmuch as better capitalized banks can be considered less risky,
these results confirm the competition-stability hypothesis. Berger, Klapper, and TurkAriss (2009) find that competition in the loan market has a mitigating effect on credit
risk. They also analyze the effect of competition on overall bank risk and banks’ capital
ratios where they find competition to increase overall bank risk and decrease capital ratios.
Finally, Schaeck and Cihak (2013 ) find in general a negative effect of competition on
bank risk for European countries. In this study we aim at adding to the empirical literature
analyzing the competition-bank risk nexus by using bank-level data for all German
banks. One of the key challenges in the attempt to identify the effect of competition and
concentration on bank risk empirically lies in defining the relevant market for each group
of banks. Existing empirical evidence generally assumes that the relevant market for a
particular bank in a given country is the country itself. This implies that each bank in a
given country stands in direct competition to all other banks in the country. While true for
large, multinational banks which compete directly with one another in many markets, this
assumption seems unrealistic for the majority of banks which operate in regional banking
markets. We improve on the existing literature by allowing competition to affect banks
operating in different markets in distinct ways. We use a total of three concepts to measure
competition, corresponding to three different dimensions at which competition might
affect the risk-taking behavior of banks. First, to approximate the ability of banks to generate
rents by pricing its products over their marginal costs, we compute bank-specific
efficiency-adjusted Lerner indexes. Second, the vast majority of banks in Germany belong
either to the cooperative banks sector or to the savings banks sector. These banks are
by law geographically limited in their scope of activities.1 Our dataset provides us with
detailed locational information for all German banks (“three-pillar system”) so that we
can exploit this special characteristic of the German banking industry to clearly define the
relevant market for each specific bank in our sample.2 For each banking market we compute
measures of concentration and contestability of the banking market. It is already a
stylized fact that measures of concentration and competition are distinct features affecting
banks in different ways. Third, we compute Boone indicators for the next contextual level
above the relevant market for banks (see also Schaeck and Cihak 2010). Although most
German banks do not run branches outside their home county, there is nothing stopping
a business customer or a depositor from choosing a bank outside that particular county.